Ontology-based Intelligent Agent for Game of Go

This paper presents a developed ontology-based intelligent agent in Monte-Carlo Tree Search (MCTS) for the computer Go application. Unlike previous research, this paper employs features derived from professional Go players´ domain knowledge to transform them into the opening-book sequence and represent them by a computer Go ontology. Afterward, the domain experts validate the built ontology. The developed computer Go ontology has been verified through the invited games for computer Go programs playing against human Go players. The results show that the ontology can work effectively for the intelligent agent and computer Go application.